diff --git a/arf.h b/arf.h
@@ -37,7 +37,7 @@
 
 #include <stdio.h>
 #include <math.h>
-#include "flint.h"
+#include "flint/flint.h"
 #include "fmpr.h"
 #include "mag.h"
 

diff --git a/arf/div.c b/arf/div.c
@@ -23,8 +23,8 @@
 
 ******************************************************************************/
 
+#include "flint/mpn_extras.h"
 #include "arf.h"
-#include "mpn_extras.h"
 
 #if !defined(__MPIR_VERSION)
 #define USE_GMP_DIV_Q 1

diff --git a/arf/get_d.c b/arf/get_d.c
@@ -23,8 +23,8 @@
 
 ******************************************************************************/
 
+#include "flint/double_extras.h"
 #include "arf.h"
-#include "double_extras.h"
 
 /* most double: (2^53-1) * 2^971 */
 /* least normal: 2^-1022 */

diff --git a/arf/set_d.c b/arf/set_d.c
@@ -23,8 +23,8 @@
 
 ******************************************************************************/
 
+#include "flint/double_extras.h"
 #include "arf.h"
-#include "double_extras.h"
 
 void
 arf_set_d(arf_t x, double v)

diff --git a/arf/test/t-add_si.c b/arf/test/t-add_si.c
@@ -24,7 +24,7 @@
 ******************************************************************************/
 
 #include "arf.h"
-#include "long_extras.h"
+#include "flint/long_extras.h"
 
 int
 arf_add_si_naive(arf_t z, const arf_t x, slong y, slong prec, arf_rnd_t rnd)

diff --git a/arf/test/t-add_ui.c b/arf/test/t-add_ui.c
@@ -24,7 +24,7 @@
 ******************************************************************************/
 
 #include "arf.h"
-#include "ulong_extras.h"
+#include "flint/ulong_extras.h"
 
 int
 arf_add_ui_naive(arf_t z, const arf_t x, ulong y, slong prec, arf_rnd_t rnd)

diff --git a/arf/test/t-addmul_si.c b/arf/test/t-addmul_si.c
@@ -24,7 +24,7 @@
 ******************************************************************************/
 
 #include "arf.h"
-#include "long_extras.h"
+#include "flint/long_extras.h"
 
 int
 arf_addmul_si_naive(arf_t z, const arf_t x, slong y, slong prec, arf_rnd_t rnd)

diff --git a/arf/test/t-mul_si.c b/arf/test/t-mul_si.c
@@ -24,7 +24,7 @@
 ******************************************************************************/
 
 #include "arf.h"
-#include "long_extras.h"
+#include "flint/long_extras.h"
 
 int
 arf_mul_si_naive(arf_t z, const arf_t x, slong y, slong prec, arf_rnd_t rnd)

diff --git a/arf/test/t-mul_ui.c b/arf/test/t-mul_ui.c
@@ -24,7 +24,7 @@
 ******************************************************************************/
 
 #include "arf.h"
-#include "ulong_extras.h"
+#include "flint/ulong_extras.h"
 
 int
 arf_mul_ui_naive(arf_t z, const arf_t x, ulong y, slong prec, arf_rnd_t rnd)

diff --git a/arf/test/t-sub_si.c b/arf/test/t-sub_si.c
@@ -24,7 +24,7 @@
 ******************************************************************************/
 
 #include "arf.h"
-#include "long_extras.h"
+#include "flint/long_extras.h"
 
 int
 arf_sub_si_naive(arf_t z, const arf_t x, slong y, slong prec, arf_rnd_t rnd)

diff --git a/arf/test/t-sub_ui.c b/arf/test/t-sub_ui.c
@@ -24,7 +24,7 @@
 ******************************************************************************/
 
 #include "arf.h"
-#include "ulong_extras.h"
+#include "flint/ulong_extras.h"
 
 int
 arf_sub_ui_naive(arf_t z, const arf_t x, ulong y, slong prec, arf_rnd_t rnd)

diff --git a/arf/test/t-submul_si.c b/arf/test/t-submul_si.c
@@ -24,7 +24,7 @@
 ******************************************************************************/
 
 #include "arf.h"
-#include "long_extras.h"
+#include "flint/long_extras.h"
 
 int
 arf_submul_si_naive(arf_t z, const arf_t x, slong y, slong prec, arf_rnd_t rnd)

diff --git a/bernoulli.h b/bernoulli.h
@@ -27,11 +27,11 @@
 #define BERNOULLI_H
 
 #include <math.h>
-#include "flint.h"
-#include "fmpz.h"
-#include "fmpz_vec.h"
-#include "fmpq.h"
-#include "arith.h"
+#include "flint/flint.h"
+#include "flint/fmpz.h"
+#include "flint/fmpz_vec.h"
+#include "flint/fmpq.h"
+#include "flint/arith.h"
 #include "arb.h"
 
 #ifdef __cplusplus

diff --git a/bernoulli/test/t-fmpq_ui.c b/bernoulli/test/t-fmpq_ui.c
@@ -25,8 +25,8 @@
 
 #include <stdio.h>
 #include <stdlib.h>
-#include "fmpz_vec.h"
-#include "arith.h"
+#include "flint/fmpz_vec.h"
+#include "flint/arith.h"
 #include "bernoulli.h"
 
 int main()

diff --git a/bernoulli/test/t-rev.c b/bernoulli/test/t-rev.c
@@ -26,9 +26,9 @@
 #include <stdio.h>
 #include <stdlib.h>
 #include "bernoulli.h"
-#include "ulong_extras.h"
-#include "nmod_poly.h"
-#include "nmod_vec.h"
+#include "flint/ulong_extras.h"
+#include "flint/nmod_poly.h"
+#include "flint/nmod_vec.h"
 
 int main()
 {

diff --git a/bool_mat.h b/bool_mat.h
@@ -33,8 +33,8 @@
 #endif
 
 #include <stdio.h>
-#include "flint.h"
-#include "fmpz_mat.h"
+#include "flint/flint.h"
+#include "flint/fmpz_mat.h"
 
 #ifdef __cplusplus
 extern "C" {

diff --git a/bool_mat/nilpotency_degree.c b/bool_mat/nilpotency_degree.c
@@ -23,8 +23,8 @@
 
 ******************************************************************************/
 
+#include "flint/fmpz_mat.h"
 #include "bool_mat.h"
-#include "fmpz_mat.h"
 
 /*
  * Postorder traversal of a DAG follows

diff --git a/bool_mat/randtest.c b/bool_mat/randtest.c
@@ -23,8 +23,8 @@
 
 ******************************************************************************/
 
+#include "flint/perm.h"
 #include "bool_mat.h"
-#include "perm.h"
 
 void
 bool_mat_randtest(bool_mat_t mat, flint_rand_t state)

diff --git a/bool_mat/test/t-all_pairs_longest_walk.c b/bool_mat/test/t-all_pairs_longest_walk.c
@@ -24,7 +24,7 @@
 ******************************************************************************/
 
 #include "bool_mat.h"
-#include "perm.h"
+#include "flint/perm.h"
 
 
 int

diff --git a/bool_mat/test/t-transitive_closure.c b/bool_mat/test/t-transitive_closure.c
@@ -24,7 +24,7 @@
 ******************************************************************************/
 
 #include "bool_mat.h"
-#include "perm.h"
+#include "flint/perm.h"
 
 /* permute rows and columns of a square matrix */
 void

diff --git a/configure b/configure
@@ -263,9 +263,16 @@ else
 fi
 
 if [ -d "${FLINT_INC_DIR}/flint" ]; then
-   FLINT_INC_DIR="${FLINT_INC_DIR}/flint"
+   FLINT_INC_DIR="${FLINT_INC_DIR}"
+elif [ -f "${FLINT_INC_DIR}/flint.h" ]; then
+   mkdir -p build/include
+   ln -sf ${FLINT_INC_DIR} build/include/flint
+   FLINT_INC_DIR="${PWD}/build/include"
 fi
 
+echo "FLINT_LIB_DIR set to ${FLINT_LIB_DIR}"
+echo "FLINT_INC_DIR set to ${FLINT_INC_DIR}"
+
 LIB_DIRS="${LIB_DIRS} ${FLINT_LIB_DIR}"
 INC_DIRS="${INC_DIRS} ${FLINT_INC_DIR}"
 LIBS="${LIBS} flint"

diff --git a/examples/complex_plot.c b/examples/complex_plot.c
@@ -4,7 +4,7 @@
 #include "acb.h"
 #include "acb_hypgeom.h"
 #include "acb_modular.h"
-#include "profiler.h"
+#include "flint/profiler.h"
 
 /* HLS algorithm from python's colorsys module */
 static double

diff --git a/examples/hilbert_matrix.c b/examples/hilbert_matrix.c
@@ -1,7 +1,7 @@
 /* This file is public domain. Author: Fredrik Johansson. */
 
 #include "arb_mat.h"
-#include "profiler.h"
+#include "flint/profiler.h"
 
 int main(int argc, char *argv[])
 {

diff --git a/examples/integrals.c b/examples/integrals.c
@@ -1,7 +1,7 @@
 /* This file is public domain. Author: Fredrik Johansson. */
 
 #include "acb_calc.h"
-#include "profiler.h"
+#include "flint/profiler.h"
 
 int
 sinx(acb_ptr out, const acb_t inp, void * params, slong order, slong prec)

diff --git a/examples/keiper_li.c b/examples/keiper_li.c
@@ -4,7 +4,7 @@
 #include "arb.h"
 #include "acb.h"
 #include "arb_poly.h"
-#include "profiler.h"
+#include "flint/profiler.h"
 
 void
 keiper_li_series(arb_ptr z, slong len, slong prec)

diff --git a/examples/pi.c b/examples/pi.c
@@ -1,7 +1,7 @@
 /* This file is public domain. Author: Fredrik Johansson. */
 
 #include "arb.h"
-#include "profiler.h"
+#include "flint/profiler.h"
 
 int main(int argc, char *argv[])
 {

diff --git a/examples/poly_roots.c b/examples/poly_roots.c
@@ -3,8 +3,8 @@
 #include <string.h>
 #include "acb.h"
 #include "acb_poly.h"
-#include "arith.h"
-#include "profiler.h"
+#include "flint/arith.h"
+#include "flint/profiler.h"
 
 int check_accuracy(acb_ptr vec, slong len, slong prec)
 {

diff --git a/examples/real_roots.c b/examples/real_roots.c
@@ -3,7 +3,7 @@
 #include <string.h>
 #include "arb_calc.h"
 #include "acb_hypgeom.h"
-#include "profiler.h"
+#include "flint/profiler.h"
 
 slong eval_count = 0;
 

diff --git a/fmpr.h b/fmpr.h
@@ -33,12 +33,12 @@
 #include <limits.h>
 #include <gmp.h>
 #include <mpfr.h>
-#include "flint.h"
-#include "fmpz.h"
-#include "fmpq.h"
+#include "flint/flint.h"
+#include "flint/fmpz.h"
+#include "flint/fmpq.h"
+#include "flint/config.h"
 #include "fmpz_extras.h"
 
-#include "config.h"
 #ifdef HAVE_TLS
 #if HAVE_TLS
 #define TLS_PREFIX __thread

diff --git a/fmpr/test/t-mul_si.c b/fmpr/test/t-mul_si.c
@@ -23,8 +23,8 @@
 
 ******************************************************************************/
 
+#include "flint/long_extras.h"
 #include "fmpr.h"
-#include "long_extras.h"
 
 static slong
 fmpr_mul_si_naive(fmpr_t z, const fmpr_t x, slong y, slong prec, fmpr_rnd_t rnd)

diff --git a/fmpr/test/t-mul_ui.c b/fmpr/test/t-mul_ui.c
@@ -23,8 +23,8 @@
 
 ******************************************************************************/
 
+#include "flint/ulong_extras.h"
 #include "fmpr.h"
-#include "ulong_extras.h"
 
 static slong
 fmpr_mul_ui_naive(fmpr_t z, const fmpr_t x, ulong y, slong prec, fmpr_rnd_t rnd)

diff --git a/fmpz_extras.h b/fmpz_extras.h
@@ -27,8 +27,8 @@
 #define FMPZ_EXTRAS_H
 
 #include <limits.h>
-#include "flint.h"
-#include "fmpz.h"
+#include "flint/flint.h"
+#include "flint/fmpz.h"
 
 #ifdef __cplusplus
 extern "C" {

diff --git a/fmpz_mat_extras.h b/fmpz_mat_extras.h
@@ -26,8 +26,8 @@
 #ifndef FMPZ_MAT_EXTRAS_H
 #define FMPZ_MAT_EXTRAS_H
 
-#include "flint.h"
-#include "fmpz_mat.h"
+#include "flint/flint.h"
+#include "flint/fmpz_mat.h"
 
 #ifdef __cplusplus
 extern "C" {

diff --git a/fmpz_mat_extras/test/t-transitive_closure.c b/fmpz_mat_extras/test/t-transitive_closure.c
@@ -24,7 +24,7 @@
 ******************************************************************************/
 
 #include "fmpz_mat_extras.h"
-#include "perm.h"
+#include "flint/perm.h"
 
 /* transitive closure can only turn zeros into ones */
 int

diff --git a/hypgeom.h b/hypgeom.h
@@ -26,9 +26,9 @@
 #ifndef HYPGEOM_H
 #define HYPGEOM_H
 
+#include "flint/fmpz_poly.h"
 #include "arb.h"
 #include "mag.h"
-#include "fmpz_poly.h"
 
 #ifdef __cplusplus
 extern "C" {

diff --git a/hypgeom/bound.c b/hypgeom/bound.c
@@ -24,7 +24,7 @@
 ******************************************************************************/
 
 #include <math.h>
-#include "double_extras.h"
+#include "flint/double_extras.h"
 #include "hypgeom.h"
 
 slong

diff --git a/hypgeom/estimate_terms_d.c b/hypgeom/estimate_terms_d.c
@@ -24,7 +24,7 @@
 ******************************************************************************/
 
 #include <math.h>
-#include "double_extras.h"
+#include "flint/double_extras.h"
 #include "hypgeom.h"
 
 #define LOG2 0.69314718055994530942

diff --git a/mag.h b/mag.h
@@ -34,8 +34,8 @@
 
 #include <stdio.h>
 #include <math.h>
-#include "flint.h"
-#include "fmpz.h"
+#include "flint/flint.h"
+#include "flint/fmpz.h"
 #include "fmpz_extras.h"
 
 #ifdef __cplusplus

diff --git a/mag/d_log.c b/mag/d_log.c
@@ -23,8 +23,8 @@
 
 ******************************************************************************/
 
+#include "flint/double_extras.h"
 #include "mag.h"
-#include "double_extras.h"
 
 /*
 This is a bad implementation the logarithm function,

diff --git a/mag/exp.c b/mag/exp.c
@@ -23,8 +23,8 @@
 
 ******************************************************************************/
 
+#include "flint/double_extras.h"
 #include "mag.h"
-#include "double_extras.h"
 
 static const double inverse_factorials[] = {
     1.0,

diff --git a/mag/get_d.c b/mag/get_d.c
@@ -23,7 +23,7 @@
 
 ******************************************************************************/
 
-#include "double_extras.h"
+#include "flint/double_extras.h"
 #include "mag.h"
 
 double

diff --git a/mag/log1p.c b/mag/log1p.c
@@ -23,8 +23,8 @@
 
 ******************************************************************************/
 
+#include "flint/double_extras.h"
 #include "mag.h"
-#include "double_extras.h"
 
 void
 mag_log1p(mag_t z, const mag_t x)

diff --git a/mag/test/t-cmp_2exp_si.c b/mag/test/t-cmp_2exp_si.c
@@ -24,7 +24,7 @@
 ******************************************************************************/
 
 #include "mag.h"
-#include "long_extras.h"
+#include "flint/long_extras.h"
 
 int main()
 {

diff --git a/mag/test/t-d_log_lower_bound.c b/mag/test/t-d_log_lower_bound.c
@@ -23,7 +23,7 @@
 
 ******************************************************************************/
 
-#include "double_extras.h"
+#include "flint/double_extras.h"
 #include "mag.h"
 
 /* XXX: d_randtest is not good enough */

diff --git a/mag/test/t-d_log_upper_bound.c b/mag/test/t-d_log_upper_bound.c
@@ -23,7 +23,7 @@
 
 ******************************************************************************/
 
-#include "double_extras.h"
+#include "flint/double_extras.h"
 #include "mag.h"
 
 /* XXX: d_randtest is not good enough */

diff --git a/mag/test/t-mul_2exp_si.c b/mag/test/t-mul_2exp_si.c
@@ -24,7 +24,7 @@
 ******************************************************************************/
 
 #include "mag.h"
-#include "long_extras.h"
+#include "flint/long_extras.h"
 
 int main()
 {

diff --git a/mag/test/t-set_d.c b/mag/test/t-set_d.c
@@ -23,7 +23,7 @@
 
 ******************************************************************************/
 
-#include "double_extras.h"
+#include "flint/double_extras.h"
 #include "mag.h"
 
 /* XXX: d_randtest is not good enough */

diff --git a/mag/test/t-set_d_2exp_fmpz.c b/mag/test/t-set_d_2exp_fmpz.c
@@ -23,7 +23,7 @@
 
 ******************************************************************************/
 
-#include "double_extras.h"
+#include "flint/double_extras.h"
 #include "mag.h"
 
 /* XXX: d_randtest is not good enough */

diff --git a/partitions.h b/partitions.h
@@ -27,8 +27,8 @@
 #define PARTITIONS_H
 
 #include <math.h>
-#include "flint.h"
-#include "arith.h"
+#include "flint/flint.h"
+#include "flint/arith.h"
 #include "arb.h"
 
 #ifdef __cplusplus

diff --git a/partitions/hrr_sum_arb.c b/partitions/hrr_sum_arb.c
@@ -24,8 +24,6 @@
 ******************************************************************************/
 
 #include "partitions.h"
-
-#include "arith.h"
 #include "arb.h"
 #include "math.h"
 

diff --git a/partitions/test/t-partitions_fmpz_ui.c b/partitions/test/t-partitions_fmpz_ui.c
@@ -23,7 +23,7 @@
 
 ******************************************************************************/
 
-#include "arith.h"
+#include "flint/arith.h"
 #include "partitions.h"
 
 /* Values mod 10^9 generated with Sage */

diff --git a/partitions/test/t-partitions_fmpz_ui_threaded.c b/partitions/test/t-partitions_fmpz_ui_threaded.c
@@ -23,7 +23,7 @@
 
 ******************************************************************************/
 
-#include "arith.h"
+#include "flint/arith.h"
 #include "partitions.h"
 
 /* Values mod 10^9 */

diff --git a/partitions/test/t-partitions_fmpz_ui_using_doubles.c b/partitions/test/t-partitions_fmpz_ui_using_doubles.c
@@ -23,7 +23,7 @@
 
 ******************************************************************************/
 
-#include "arith.h"
+#include "flint/arith.h"
 #include "partitions.h"
 
 /* Values mod 10^9 generated with Sage */